The hemiparetic patients, who had a moderate amount of functional recovery, were similar to healthy subjects in their ability to control reach-to-grasp components. Their performance was not as skilled, however.

Programming and execution of arm movements in Parkinson's disease were investigated in choice and simple reaction time (RT) situations in which subjects made aimed movements at a target. A no-aiming condition was also studied. Reaction time was fractionated using surface EMG recording into premotor (central) and motor (peripheral) components. Premotor RT was found to be greater for parkinsonian patients than normal age-matched controls in the simple RT condition, but not in the choice condition. This effect did not depend on the parameters of the impending movement. Thus, paradoxically, parkinsonian patients were not inherently slower at initiating aiming movements from the starting position, but seemed unable to use advance information concerning motor task demands to speed up movement initiation. For both groups, low velocity movements took longer to initiate than high velocity ones. In the no-aiming condition parkinsonian RTs were markedly shorter than when aiming, but were still significantly longer than control RTs. Motor RT was constant across all conditions and was not different for patient and control subjects. In all conditions, parkinsonian movements were around 37% slower than control movements, and their movement times were more variable, the differences showing up early on in the movement, that is, during the initial ballistic phase. The within-subject variability of movement endpoints was also greater in patients. The motor dysfunction displayed in Parkinson's disease involves a number of components: (1) a basic central problem with simply initiating movements, even when minimal programming is required (no-aiming condition); (2) difficulty in maintaining computed forces for motor programs over time (simple RT condition); (3) a basic slowness of movement (bradykinesia) in all conditions; and (4) increased variability of movement in both time and space, presumably caused by inherent variability in force production.

When making discrete aiming movements, patients with Parkinson's disease show greater inherent variability in the endpoints of their movements than do normal subjects. Endpoint variability can be reduced, by moving more slowly, by utilizing visual guidance, and by making small amplitude movements. The greater variability of patients is not a universal finding, but depends on the conditions of movement. For small movements the performance of patients equates to that of controls. For larger movements the results indicate that if sufficient time is available, patients can use visual guidance (if available) to reduce the variability of their movements to the level of normals. Patients can generate fast and/or large amplitude arm movements if required, but they are erratic if made in the dark or over a short duration. Their difficulty lies not so much in the magnitude of muscle force available to them, but rather in an inability to produce it consistently for any given movement attempted. Bradykinesia may in part result from this inherent variability in that parkinsonian patients, in order to maintain accuracy within acceptable limits, are forced to increase the duration of their movements to a level where they can make use of visual guidance. In any event, theoretical explanations for the movement disorder in Parkinson's disease advanced in the literature need to take some account of this increased variability of movement.

No abstract

This paper presents a reappraisal of Fitts' (1954) law and outlines reasoning which predicts that, for tasks of equivalent difficulty as calculated by Fitts' Index of Difficulty, smaller-tolerance tasks will, within limits, take longer to perform. Reanalysis of Fitts' data appears to confirm this view. The notion that the overall performance of the human motor system plus associated feedback mechanisms can be described by any simple formulation assuming constant weighting of factors over a variety of tasks, such as proposed by Fitts' law, seems to be misfounded. Recent research supports the doubts concerning fixed-parameter linear and nonlinear models of the human motor system. Some implications of these ideas are discussed.